The invention relates to a method of manufacturing a cathode ray tube, which method comprises a step of inspecting a pattern provided on a display screen of or for a cathode ray tube.
Cathode ray tubes are used, inter alia, for (color) display devices such as televisions and computer monitors.
A cathode ray tube comprises a display screen which has a plurality of patterns, for example, phosphor patterns or black matrix patterns. Errors in these patterns are visible in the image displayed by the cathode ray tube and may thus have the result that the display device does not satisfy the imposed quality requirements and is consequently unsalable and has to be rejected. It is therefore important to detect pattern errors at an early stage of manufacture in order to reduce the number of rejects and/or to be able to interfere with the manufacturing process at an early stage so as to reduce the number of errors. A method of the type described in the opening paragraph is known from the English-language abstract of Japanese patent application JP 53-48589. This abstract describes a method in which a movable mirror is arranged in the focus of a spherical display window (a display window whose shape can be described by one radius of curvature) of or for a cathode ray tube, which movable mirror images pixels of the frame line by line in a small aperture behind which a light detector is arranged. The data picked up by the light detector are applied to a processing unit.
Although inspection of a display screen is possible in this manner, this method has some shortcomings.
The method may be applicable for inspecting screens at random, but not in a production line or in large numbers. The reason is that this requires a long pick-up and processing time for all data of a display screen. Moreover, problems occur with regard to display screens having an aspherical shape or display screens having a large radius of curvature, because the distance between the display window and the mirror becomes large and thus the arrangement becomes large and occupies much space. As the distance between the display screen and optical elements such as mirrors and lenses becomes larger, the resolution of the arrangement (=the smallest detail which can still be distinguished) becomes smaller. This can be corrected by using larger mirrors or lenses but this has the drawback that the mass of the elements increases so that the inertia increases, making it more difficult to manufacture optical elements of high quality.
It is an object of the invention to provide a method of the type described in the opening paragraph, in which one or more of said problems or problems mentioned below is obviated.
To this end, the method according to the invention is characterized in that a set of sub-images of the display screen, which set preferably covers the display screen entirely, is guided towards a camera via an adjustable mirror arrangement and an adjustable focal length lens (hereinafter adjustable focal lens or focal lens), each sub-image being picked up with the mirror arrangement and the focal lens in a rest state, the data of each sub-image being applied to a processing unit in which difference data about differences between the picked-up image and a standard image are generated, said data being applied to a further processing unit.
By using an adjustable focal lens, each sub-image can be sharply imaged on pick-up element of the camera. Changes in the distance between the camera and the display screen can thus be compensated. The mirror arrangement and the focal lens is set to the envisaged position for each sub-image. Subsequently, the sub-image, thus with the mirror and the focal lens in the rest (non-moving) state is picked up.
A problem in picking up images is that vibrations which may result from moving parts can disturb the pick-up operation. A xe2x80x9cflying-spotxe2x80x9d method, i.e. a method in which measurements are performed while elements (such as mirrors) are moving, thus disturbs the pick-up operation. Such disturbances may result in xe2x80x9cfalsexe2x80x9d errors, i.e. errors are registered as a result of movements of elements of the measuring arrangements. xe2x80x9cFalsexe2x80x9d errors raise the number of rejects and may cause considerable damage. The occurrence of vibrations may be prevented, for example, by maintaining the speed at which elements are moved relatively low, but this means that the measuring time is relatively long. The consequences of vibrations may also be decreased, for example, by damping measures, but this has a cost-increasing effect.
In the method according to the invention, the movable elements such as the mirror arrangement and the focal lens do not move when the sub-images are being picked up. As compared with the known method, in which the mirror arrangement is placed in the focus of the display window, the method according to the invention provides the additional advantage that the distance between the mirror arrangement and the display screen can be reduced as a result of the use of the adjustable focal lens. Nowadays, display windows are used with radii of curvature of several meters (2-5 m). In the method according to the invention, the distance between the display screen and the mirror arrangement is preferably between 50 cm and 120 cm so that a considerably shorter distance between the display screen and the mirror arrangement is possible. Thus the ratio of screen radius of curvature to mirror-to-screen distance may be greater than 16:1. This shorter distance has two positive consequences.
The method occupies less space and the effects of vibration of the mirror arrangement on an image picked up by the camera are smaller.
The distance is preferably shorter than 120 cm. As the distance increases, the resolution decreases. The resolution is approximately proportional to the cross-section of the optical elements, divided by the square value of the distance. Relatively small optical elements may be used up to distances of approximately 120 cm.
The distance is preferably larger than 50 cm. It is true that if only distance and size of the optical elements are taken into consideration, the resolution increases as the distance decreases, but this is offset by the fact that as the distance decreases, the angle at which sub-images of the edges of the display screen are picked up increases more and more. This increase of the angle causes a distortion of the image, and defocus of the image because not every pixel of the image can be sharply focused. These effects also lead to a decrease of the resolution.
Preferably, the sub-images at least partly overlap each other. Sub-image overlaps have the advantage that errors can be satisfactorily localized.
Preferably, between 100 (for example, 10xc3x9710) and 1200 (for example, 40xc3x9730) sub-images are picked up per display screen. The data of each sub-image are applied to a processing unit. The time required by this processing unit to process the data and generate differences between the picked up image and the standard image (hereinafter also referred to as xe2x80x9cprocessing timexe2x80x9d) should be of the order of the time which is required to adjust the mirror arrangement and the focal lens and to dampen vibrations (hereinafter also referred to as xe2x80x9cadjusting timexe2x80x9d). The processing time increases as the number of sub-images decreases, so that the surface of a sub-image increases. This is, however, no linear relation, but the processing time increases faster than linearly with the surface. From a point of view of processing the data, it is favorable to pick up small and many sub-images. However, a waiting time is to be observed between each pick-up operation in order to adjust the mirror arrangement and the focal lens and to dampen vibrations. It is not possible to measure during an adjusting time. For reasons of efficiency, it is therefore favorable to maintain the total adjusting time (which is equal to the number of sub-images multiplied by an adjusting time) to be short and thus to pick up a small number of images. This seems to be a dilemma. However, the inventors have realized that, if the processing time is approximately equal to or several times larger than the adjusting time, the data of a sub-image can be entirely processed or processed for the greater part while the mirror arrangement and the focal lens are being adjusted for the next sub-image, thus resulting in a very efficient method.
Except for an initial sub-image and a final sub-image, an adjacent sub-image is preferably picked up after each sub-image. Consequently, the display screen is preferably not scanned from left to right and line by line, as is common practice with electron beams in cathode ray tubes, but each sub-image (with the exception of an initial sub-image and a final sub-image) adjoins and preferably partially overlaps the previous sub-image. The optical elements (such as mirror arrangement and focal lens) thus make small movements so that the adjusting time is short and, moreover, the adjusting time between two sub-images is approximately equal for each sub-image. The total adjusting time is thus relatively short. Moreover, a comparison of sub-images is simpler and difference data can therefore be generated more rapidly.
In the method according to the invention, difference data about differences between the picked up image and the standard image are generated in the processing unit, which data are applied to a further processing unit. Since analysis takes place per sub-image and all important (difference) data are applied to the further processing unit, the processing unit needs to be provided with only a relatively small memory (as compared with a method in which the total image is picked up in one operation). This has a cost-decreasing effect and raises the speed with which measurements can take place.
The processing unit only applies those difference data which exceed a threshold value to the further processing unit.
The processing unit may apply all data to the further processing unit. However, the inventors have realized that it is more efficient to filter the difference data before they are applied to the further processing unit and to apply only differences which are larger than a threshold value. Such a filter considerably reduces the number of data to be processed.
In the further processing unit, the difference data are preferably processed, ordered as to magnitude and/or importance, and these ordered data are applied to a display device for the purpose of display. The further processing unit processes the difference data to errors in the image and orders these data as to magnitude and/or importance. The result is displayed in a display device. By displaying the image errors as to magnitude and/or importance, an operator will be able to judge these errors. In the current state of the art, this is a great advantage. Not every xe2x80x9cerrorxe2x80x9d which is displayed is necessarily a real error. A method in which inspection of the detected errors by a human being is possible thus has great advantages. The further processing unit preferably comprises a means for introducing data about errors in the image. This provides the possibility of supplying data about errors in the image to the further processing unit. This may be used, for example, for suppressing frequently recurrent xe2x80x9cfalsexe2x80x9d errors.
These and other aspects of the invention are apparent from and will be elucidated with reference to the embodiments described hereinafter.